Thesis
Complex mixed-methods research design approaches for data mining to meet utility and trust requirements for three-phase energy end-use customers
- Creator
- Rights statement
- Awarding institution
- University of Strathclyde
- Date of award
- 2025
- Thesis identifier
- T17340
- Person Identifier (Local)
- 202163186
- Qualification Level
- Qualification Name
- Department, School or Faculty
- Abstract
- This thesis explores complex mixed-methods research design approaches for data mining to meet utility and trust requirements for three-phase energy end-use customers (residential and industrial) of electricity networks. The developed methods explore how much information can be inferred about individual load usage from single- and three-phase installations through Non-Intrusive Load Monitoring (NILM) using granular metered data, from low-frequency (1-sec) power data obtained through special metering equipment to very-low frequency (30-min) energy data obtained through the available smart metering infrastructure. Generalisability, i.e., training and testing in different premises that exhibit similar load patterns, and transferability, i.e., training and testing in different premises that exhibit different load patterns, has been explored in order to co-design scalable downstream NILM applications with end-users. Quantitative data, i.e., metered physical quantities, are complemented by qualitative data obtained from interviews and time-of-use surveys in order to develop complex mixed-methods data mining approaches to inform: (i) the usage component of lifecycle assessment (LCA) models of electric vehicles (EVs); (ii) the evaluation of energy-efficiency in net-positive energy households; and, (iii) load-scheduling of energy intensive activities for the residential and industrial sectors. Therefore, the goals of the thesis can be summarised as follows: (i) explore the effects of different levels of smart metering data granularity on the load disaggregation accuracy and robustness; (ii) improve the load disaggregation accuracy when using very-low frequency data obtained through the smart metering infrastructure by exploiting three-phase information in residential and industrial settings; (iii) generalise and transfer the trained models across different settings at scale; (iv) inform the usage component of LCA models of EVs and compare them fairly with the fossil-fuelled equivalents, through the disaggregation of EV charging loads in residential customers and integration of drivers’ charging routines by combining quantitative — i.e., energy consumption and production timeseries and granular spatio-temporal carbon intensity of the electricity network — and qualitative data including interviews and time-of-use surveys; (v) develop complex mixed-methods data-driven energy centric evaluation methods of net-positive households’ methodology to answer the “what”, “why” and “how” of energy prosumption in net-positive energy neighbourhoods through the disaggregation of energy-intensive load to the activity level and explore the potential of load scheduling on an activity basis based on different households profiles and flexibilities; and, (vi) develop a co-created NILM-enabled data driven methodology to improve load scheduling in the dairy sector and reduce the utility costs and the carbon footprint of farms, through the collaboration with various stakeholders during the design, data collection, implementation, and feedback process. By achieving these goals, this thesis aims to address all three levels (biosphere, society, and economy) of the Sustainable Development Goals (SDGs). By enhancing LCA of EVs, national policies can be directly influenced (SDG 13.2), while at the same time reducing pollution for sustainable cities (SDG 11.6), ensuring access to clean energy (SDG 7.1), and promoting responsible electricity consumption (SDG 12.2). The mixed methods approach for net-positive buildings supports climate action policies (SDG 13.2), sustainable urbanisation (SDG 11.3), clean energy access (SDG 7.1), and efficient resource use (SDG 12.2). In the agricultural sector, research improves renewable energy generation, self-consumption, and sustainability (SDGs 7.2, 11.5, 12.2, 13.2), enhances resource efficiency and clean technology adoption (SDG 8.4, 9.4), and reduces inequalities by supporting income growth in Less Favoured Areas (SDG 10.1).
- Advisor / supervisor
- Stanković, Vladimir
- Stanković, Lina
- Resource Type
- DOI
Las relaciones
Elementos
| Miniatura | Título | Fecha de subida | Visibilidad | Acciones |
|---|---|---|---|---|
|
|
PDF of thesis T17340 | 2025-07-03 | Público | Descargar |